intl/dcgettext.c

   1 /* Implementation of the dcgettext(3) function.
   2    Copyright (C) 1995, 1996, 1997 Free Software Foundation, Inc.
   3
   4    This file is part of the GNU C Library.  Its master source is NOT part of
   5    the C library, however.
   6
   7    The GNU C Library is free software; you can redistribute it and/or
   8    modify it under the terms of the GNU Library General Public License as
   9    published by the Free Software Foundation; either version 2 of the
  10    License, or (at your option) any later version.
  11
  12    The GNU C Library is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15    Library General Public License for more details.
  16
  17    You should have received a copy of the GNU Library General Public
  18    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  19    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  20    Boston, MA 02111-1307, USA.  */
  21
  22 #ifdef HAVE_CONFIG_H
  23 # include <config.h>
  24 #endif
  25
  26 #include <sys/types.h>
  27
  28 #ifdef __GNUC__
  29 # define alloca __builtin_alloca
  30 # define HAVE_ALLOCA 1
  31 #else
  32 # if defined HAVE_ALLOCA_H || defined _LIBC
  33 #  include <alloca.h>
  34 # else
  35 #  ifdef _AIX
  36  #pragma alloca
  37 #  else
  38 #   ifndef alloca
  39 char *alloca ();
  40 #   endif
  41 #  endif
  42 # endif
  43 #endif
  44
  45 #include <errno.h>
  46 #ifndef errno
  47 extern int errno;
  48 #endif
  49 #ifndef __set_errno
  50 # define __set_errno(val) errno = (val)
  51 #endif
  52
  53 #if defined STDC_HEADERS || defined _LIBC
  54 # include <stdlib.h>
  55 #else
  56 char *getenv ();
  57 # ifdef HAVE_MALLOC_H
  58 #  include <malloc.h>
  59 # else
  60 void free ();
  61 # endif
  62 #endif
  63
  64 #if defined HAVE_STRING_H || defined _LIBC
  65 # ifndef _GNU_SOURCE
  66 #  define _GNU_SOURCE   1
  67 # endif
  68 # include <string.h>
  69 #else
  70 # include <strings.h>
  71 #endif
  72 #if !HAVE_STRCHR && !defined _LIBC
  73 # ifndef strchr
  74 #  define strchr index
  75 # endif
  76 #endif
  77
  78 #if defined HAVE_UNISTD_H || defined _LIBC
  79 # include <unistd.h>
  80 #endif
  81
  82 #include "gettext.h"
  83 #include "gettextP.h"
  84 #ifdef _LIBC
  85 # include <libintl.h>
  86 #else
  87 # include "libgettext.h"
  88 #endif
  89 #include "hash-string.h"
  90
  91 /* @@ end of prolog @@ */
  92
  93 #ifdef _LIBC
  94 /* Rename the non ANSI C functions.  This is required by the standard
  95    because some ANSI C functions will require linking with this object
  96    file and the name space must not be polluted.  */
  97 # define getcwd __getcwd
  98 # ifndef stpcpy
  99 #  define stpcpy __stpcpy
 100 # endif
 101 #else
 102 # if !defined HAVE_GETCWD
 103 char *getwd ();
 104 #  define getcwd(buf, max) getwd (buf)
 105 # else
 106 char *getcwd ();
 107 # endif
 108 # ifndef HAVE_STPCPY
 109 static char *stpcpy PARAMS ((char *dest, const char *src));
 110 # endif
 111 #endif
 112
 113 /* Amount to increase buffer size by in each try.  */
 114 #define PATH_INCR 32
 115
 116 /* The following is from pathmax.h.  */
 117 /* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define
 118    PATH_MAX but might cause redefinition warnings when sys/param.h is
 119    later included (as on MORE/BSD 4.3).  */
 120 #if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && !defined(__GNUC__))
 121 # include <limits.h>
 122 #endif
 123
 124 #ifndef _POSIX_PATH_MAX
 125 # define _POSIX_PATH_MAX 255
 126 #endif
 127
 128 #if !defined(PATH_MAX) && defined(_PC_PATH_MAX)
 129 # define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX))
 130 #endif
 131
 132 /* Don't include sys/param.h if it already has been.  */
 133 #if defined(HAVE_SYS_PARAM_H) && !defined(PATH_MAX) && !defined(MAXPATHLEN)
 134 # include <sys/param.h>
 135 #endif
 136
 137 #if !defined(PATH_MAX) && defined(MAXPATHLEN)
 138 # define PATH_MAX MAXPATHLEN
 139 #endif
 140
 141 #ifndef PATH_MAX
 142 # define PATH_MAX _POSIX_PATH_MAX
 143 #endif
 144
 145 #ifndef internal_function
 146 # define internal_function
 147 #endif
 148
 149 /* XPG3 defines the result of `setlocale (category, NULL)' as:
 150    ``Directs `setlocale()' to query `category' and return the current
 151      setting of `local'.''
 152    However it does not specify the exact format.  And even worse: POSIX
 153    defines this not at all.  So we can use this feature only on selected
 154    system (e.g. those using GNU C Library).  */
 155 #ifdef _LIBC
 156 # define HAVE_LOCALE_NULL
 157 #endif
 158
 159 /* Name of the default domain used for gettext(3) prior any call to
 160    textdomain(3).  The default value for this is "messages".  */
 161 const char _nl_default_default_domain[] = "messages";
 162
 163 /* Value used as the default domain for gettext(3).  */
 164 const char *_nl_current_default_domain = _nl_default_default_domain;
 165
 166 /* Contains the default location of the message catalogs.  */
 167 const char _nl_default_dirname[] = GNULOCALEDIR;
 168
 169 /* List with bindings of specific domains created by bindtextdomain()
 170    calls.  */
 171 struct binding *_nl_domain_bindings;
 172
 173 /* Prototypes for local functions.  */
 174 static char *find_msg PARAMS ((struct loaded_l10nfile *domain_file,
 175                                const char *msgid)) internal_function;
 176 static const char *category_to_name PARAMS ((int category)) internal_function;
 177 static const char *guess_category_value PARAMS ((int category,
 178                                                  const char *categoryname))
 179      internal_function;
 180
 181
 182 /* For those loosing systems which don't have `alloca' we have to add
 183    some additional code emulating it.  */
 184 #ifdef HAVE_ALLOCA
 185 /* Nothing has to be done.  */
 186 # define ADD_BLOCK(list, address) /* nothing */
 187 # define FREE_BLOCKS(list) /* nothing */
 188 #else
 189 struct block_list
 190 {
 191   void *address;
 192   struct block_list *next;
 193 };
 194 # define ADD_BLOCK(list, addr)                                                \
 195   do {                                                                        \
 196     struct block_list *newp = (struct block_list *) malloc (sizeof (*newp));  \
 197     /* If we cannot get a free block we cannot add the new element to         \
 198        the list.  */                                                          \
 199     if (newp != NULL) {                                                       \
 200       newp->address = (addr);                                                 \
 201       newp->next = (list);                                                    \
 202       (list) = newp;                                                          \
 203     }                                                                         \
 204   } while (0)
 205 # define FREE_BLOCKS(list)                                                    \
 206   do {                                                                        \
 207     while (list != NULL) {                                                    \
 208       struct block_list *old = list;                                          \
 209       list = list->next;                                                      \
 210       free (old);                                                             \
 211     }                                                                         \
 212   } while (0)
 213 # undef alloca
 214 # define alloca(size) (malloc (size))
 215 #endif  /* have alloca */
 216
 217
 218 /* Names for the libintl functions are a problem.  They must not clash
 219    with existing names and they should follow ANSI C.  But this source
 220    code is also used in GNU C Library where the names have a __
 221    prefix.  So we have to make a difference here.  */
 222 #ifdef _LIBC
 223 # define DCGETTEXT __dcgettext
 224 #else
 225 # define DCGETTEXT dcgettext__
 226 #endif
 227
 228 /* Look up MSGID in the DOMAINNAME message catalog for the current CATEGORY
 229    locale.  */
 230 char *
 231 DCGETTEXT (domainname, msgid, category)
 232      const char *domainname;
 233      const char *msgid;
 234      int category;
 235 {
 236 #ifndef HAVE_ALLOCA
 237   struct block_list *block_list = NULL;
 238 #endif
 239   struct loaded_l10nfile *domain;
 240   struct binding *binding;
 241   const char *categoryname;
 242   const char *categoryvalue;
 243   char *dirname, *xdomainname;
 244   char *single_locale;
 245   char *retval;
 246   int saved_errno = errno;
 247
 248   /* If no real MSGID is given return NULL.  */
 249   if (msgid == NULL)
 250     return NULL;
 251
 252   /* If DOMAINNAME is NULL, we are interested in the default domain.  If
 253      CATEGORY is not LC_MESSAGES this might not make much sense but the
 254      defintion left this undefined.  */
 255   if (domainname == NULL)
 256     domainname = _nl_current_default_domain;
 257
 258   /* First find matching binding.  */
 259   for (binding = _nl_domain_bindings; binding != NULL; binding = binding->next)
 260     {
 261       int compare = strcmp (domainname, binding->domainname);
 262       if (compare == 0)
 263         /* We found it!  */
 264         break;
 265       if (compare < 0)
 266         {
 267           /* It is not in the list.  */
 268           binding = NULL;
 269           break;
 270         }
 271     }
 272
 273   if (binding == NULL)
 274     dirname = (char *) _nl_default_dirname;
 275   else if (binding->dirname[0] == '/')
 276     dirname = binding->dirname;
 277   else
 278     {
 279       /* We have a relative path.  Make it absolute now.  */
 280       size_t dirname_len = strlen (binding->dirname) + 1;
 281       size_t path_max;
 282       char *ret;
 283
 284       path_max = (unsigned) PATH_MAX;
 285       path_max += 2;            /* The getcwd docs say to do this.  */
 286
 287       dirname = (char *) alloca (path_max + dirname_len);
 288       ADD_BLOCK (block_list, dirname);
 289
 290       __set_errno (0);
 291       while ((ret = getcwd (dirname, path_max)) == NULL && errno == ERANGE)
 292         {
 293           path_max += PATH_INCR;
 294           dirname = (char *) alloca (path_max + dirname_len);
 295           ADD_BLOCK (block_list, dirname);
 296           __set_errno (0);
 297         }
 298
 299       if (ret == NULL)
 300         {
 301           /* We cannot get the current working directory.  Don't signal an
 302              error but simply return the default string.  */
 303           FREE_BLOCKS (block_list);
 304           __set_errno (saved_errno);
 305           return (char *) msgid;
 306         }
 307
 308       stpcpy (stpcpy (strchr (dirname, '\0'), "/"), binding->dirname);
 309     }
 310
 311   /* Now determine the symbolic name of CATEGORY and its value.  */
 312   categoryname = category_to_name (category);
 313   categoryvalue = guess_category_value (category, categoryname);
 314
 315   xdomainname = (char *) alloca (strlen (categoryname)
 316                                  + strlen (domainname) + 5);
 317   ADD_BLOCK (block_list, xdomainname);
 318
 319   stpcpy (stpcpy (stpcpy (stpcpy (xdomainname, categoryname), "/"),
 320                   domainname),
 321           ".mo");
 322
 323   /* Creating working area.  */
 324   single_locale = (char *) alloca (strlen (categoryvalue) + 1);
 325   ADD_BLOCK (block_list, single_locale);
 326
 327
 328   /* Search for the given string.  This is a loop because we perhaps
 329      got an ordered list of languages to consider for th translation.  */
 330   while (1)
 331     {
 332       /* Make CATEGORYVALUE point to the next element of the list.  */
 333       while (categoryvalue[0] != '\0' && categoryvalue[0] == ':')
 334         ++categoryvalue;
 335       if (categoryvalue[0] == '\0')
 336         {
 337           /* The whole contents of CATEGORYVALUE has been searched but
 338              no valid entry has been found.  We solve this situation
 339              by implicitly appending a "C" entry, i.e. no translation
 340              will take place.  */
 341           single_locale[0] = 'C';
 342           single_locale[1] = '\0';
 343         }
 344       else
 345         {
 346           char *cp = single_locale;
 347           while (categoryvalue[0] != '\0' && categoryvalue[0] != ':')
 348             *cp++ = *categoryvalue++;
 349           *cp = '\0';
 350         }
 351
 352       /* If the current locale value is C (or POSIX) we don't load a
 353          domain.  Return the MSGID.  */
 354       if (strcmp (single_locale, "C") == 0
 355           || strcmp (single_locale, "POSIX") == 0)
 356         {
 357           FREE_BLOCKS (block_list);
 358           __set_errno (saved_errno);
 359           return (char *) msgid;
 360         }
 361
 362
 363       /* Find structure describing the message catalog matching the
 364          DOMAINNAME and CATEGORY.  */
 365       domain = _nl_find_domain (dirname, single_locale, xdomainname);
 366
 367       if (domain != NULL)
 368         {
 369           retval = find_msg (domain, msgid);
 370
 371           if (retval == NULL)
 372             {
 373               int cnt;
 374
 375               for (cnt = 0; domain->successor[cnt] != NULL; ++cnt)
 376                 {
 377                   retval = find_msg (domain->successor[cnt], msgid);
 378
 379                   if (retval != NULL)
 380                     break;
 381                 }
 382             }
 383
 384           if (retval != NULL)
 385             {
 386               FREE_BLOCKS (block_list);
 387               __set_errno (saved_errno);
 388               return retval;
 389             }
 390         }
 391     }
 392   /* NOTREACHED */
 393 }
 394
 395 #ifdef _LIBC
 396 /* Alias for function name in GNU C Library.  */
 397 weak_alias (__dcgettext, dcgettext);
 398 #endif
 399
 400
 401 static char *
 402 internal_function
 403 find_msg (domain_file, msgid)
 404      struct loaded_l10nfile *domain_file;
 405      const char *msgid;
 406 {
 407   size_t top, act, bottom;
 408   struct loaded_domain *domain;
 409
 410   if (domain_file->decided == 0)
 411     _nl_load_domain (domain_file);
 412
 413   if (domain_file->data == NULL)
 414     return NULL;
 415
 416   domain = (struct loaded_domain *) domain_file->data;
 417
 418   /* Locate the MSGID and its translation.  */
 419   if (domain->hash_size > 2 && domain->hash_tab != NULL)
 420     {
 421       /* Use the hashing table.  */
 422       nls_uint32 len = strlen (msgid);
 423       nls_uint32 hash_val = hash_string (msgid);
 424       nls_uint32 idx = hash_val % domain->hash_size;
 425       nls_uint32 incr = 1 + (hash_val % (domain->hash_size - 2));
 426       nls_uint32 nstr = W (domain->must_swap, domain->hash_tab[idx]);
 427
 428       if (nstr == 0)
 429         /* Hash table entry is empty.  */
 430         return NULL;
 431
 432       if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 433           && strcmp (msgid,
 434                      domain->data + W (domain->must_swap,
 435                                        domain->orig_tab[nstr - 1].offset)) == 0)
 436         return (char *) domain->data + W (domain->must_swap,
 437                                           domain->trans_tab[nstr - 1].offset);
 438
 439       while (1)
 440         {
 441           if (idx >= domain->hash_size - incr)
 442             idx -= domain->hash_size - incr;
 443           else
 444             idx += incr;
 445
 446           nstr = W (domain->must_swap, domain->hash_tab[idx]);
 447           if (nstr == 0)
 448             /* Hash table entry is empty.  */
 449             return NULL;
 450
 451           if (W (domain->must_swap, domain->orig_tab[nstr - 1].length) == len
 452               && strcmp (msgid,
 453                          domain->data + W (domain->must_swap,
 454                                            domain->orig_tab[nstr - 1].offset))
 455                  == 0)
 456             return (char *) domain->data
 457               + W (domain->must_swap, domain->trans_tab[nstr - 1].offset);
 458         }
 459       /* NOTREACHED */
 460     }
 461
 462   /* Now we try the default method:  binary search in the sorted
 463      array of messages.  */
 464   bottom = 0;
 465   top = domain->nstrings;
 466   while (bottom < top)
 467     {
 468       int cmp_val;
 469
 470       act = (bottom + top) / 2;
 471       cmp_val = strcmp (msgid, domain->data
 472                                + W (domain->must_swap,
 473                                     domain->orig_tab[act].offset));
 474       if (cmp_val < 0)
 475         top = act;
 476       else if (cmp_val > 0)
 477         bottom = act + 1;
 478       else
 479         break;
 480     }
 481
 482   /* If an translation is found return this.  */
 483   return bottom >= top ? NULL : (char *) domain->data
 484                                 + W (domain->must_swap,
 485                                      domain->trans_tab[act].offset);
 486 }
 487
 488
 489 /* Return string representation of locale CATEGORY.  */
 490 static const char *
 491 internal_function
 492 category_to_name (category)
 493      int category;
 494 {
 495   const char *retval;
 496
 497   switch (category)
 498   {
 499 #ifdef LC_COLLATE
 500   case LC_COLLATE:
 501     retval = "LC_COLLATE";
 502     break;
 503 #endif
 504 #ifdef LC_CTYPE
 505   case LC_CTYPE:
 506     retval = "LC_CTYPE";
 507     break;
 508 #endif
 509 #ifdef LC_MONETARY
 510   case LC_MONETARY:
 511     retval = "LC_MONETARY";
 512     break;
 513 #endif
 514 #ifdef LC_NUMERIC
 515   case LC_NUMERIC:
 516     retval = "LC_NUMERIC";
 517     break;
 518 #endif
 519 #ifdef LC_TIME
 520   case LC_TIME:
 521     retval = "LC_TIME";
 522     break;
 523 #endif
 524 #ifdef LC_MESSAGES
 525   case LC_MESSAGES:
 526     retval = "LC_MESSAGES";
 527     break;
 528 #endif
 529 #ifdef LC_RESPONSE
 530   case LC_RESPONSE:
 531     retval = "LC_RESPONSE";
 532     break;
 533 #endif
 534 #ifdef LC_ALL
 535   case LC_ALL:
 536     /* This might not make sense but is perhaps better than any other
 537        value.  */
 538     retval = "LC_ALL";
 539     break;
 540 #endif
 541   default:
 542     /* If you have a better idea for a default value let me know.  */
 543     retval = "LC_XXX";
 544   }
 545
 546   return retval;
 547 }
 548
 549 /* Guess value of current locale from value of the environment variables.  */
 550 static const char *
 551 internal_function
 552 guess_category_value (category, categoryname)
 553      int category;
 554      const char *categoryname;
 555 {
 556   const char *retval;
 557
 558   /* The highest priority value is the `LANGUAGE' environment
 559      variable.  This is a GNU extension.  */
 560   retval = getenv ("LANGUAGE");
 561   if (retval != NULL && retval[0] != '\0')
 562     return retval;
 563
 564   /* `LANGUAGE' is not set.  So we have to proceed with the POSIX
 565      methods of looking to `LC_ALL', `LC_xxx', and `LANG'.  On some
 566      systems this can be done by the `setlocale' function itself.  */
 567 #if defined HAVE_SETLOCALE && defined HAVE_LC_MESSAGES && defined HAVE_LOCALE_NULL
 568   return setlocale (category, NULL);
 569 #else
 570   /* Setting of LC_ALL overwrites all other.  */
 571   retval = getenv ("LC_ALL");
 572   if (retval != NULL && retval[0] != '\0')
 573     return retval;
 574
 575   /* Next comes the name of the desired category.  */
 576   retval = getenv (categoryname);
 577   if (retval != NULL && retval[0] != '\0')
 578     return retval;
 579
 580   /* Last possibility is the LANG environment variable.  */
 581   retval = getenv ("LANG");
 582   if (retval != NULL && retval[0] != '\0')
 583     return retval;
 584
 585   /* We use C as the default domain.  POSIX says this is implementation
 586      defined.  */
 587   return "C";
 588 #endif
 589 }
 590
 591 /* @@ begin of epilog @@ */
 592
 593 /* We don't want libintl.a to depend on any other library.  So we
 594    avoid the non-standard function stpcpy.  In GNU C Library this
 595    function is available, though.  Also allow the symbol HAVE_STPCPY
 596    to be defined.  */
 597 #if !_LIBC && !HAVE_STPCPY
 598 static char *
 599 stpcpy (dest, src)
 600      char *dest;
 601      const char *src;
 602 {
 603   while ((*dest++ = *src++) != '\0')
 604     /* Do nothing. */ ;
 605   return dest - 1;
 606 }
 607 #endif
 608
 609
 610 #ifdef _LIBC
 611 /* If we want to free all resources we have to do some work at
 612    program's end.  */
 613 static void __attribute__ ((unused))
 614 free_mem (void)
 615 {
 616   struct binding *runp;
 617
 618   for (runp = _nl_domain_bindings; runp != NULL; runp = runp->next)
 619     {
 620       free (runp->domainname);
 621       if (runp->dirname != _nl_default_dirname)
 622         /* Yes, this is a pointer comparison.  */
 623         free (runp->dirname);
 624     }
 625
 626   if (_nl_current_default_domain != _nl_default_default_domain)
 627     /* Yes, again a pointer comparison.  */
 628     free ((char *) _nl_current_default_domain);
 629 }
 630
 631 text_set_element (__libc_subfreeres, free_mem);
 632 #endif